Hagstrom SA, Ying G, Pauer GJT, et al; Comparison of Age-
Related Macular Degeneration Treatments Trials (CATT) Research
Group. VEGFA and VEGFR2 gene polymorphisms and response to anti –vascular endothelial growth factor therapy: Comparison of
Age-Related Macular Degeneration Treatments Trials (CATT).
JAMA Ophthalmol . Published online March 20, 2014. doi:10.1001/jamaophthalmol.2014.109.
Copyright restrictions may apply

Importance
• Individual variation in response and duration of anti –vascular endothelial growth factor (VEGF) therapy is seen among patients with neovascular age-related macular degeneration (nAMD). Identification of genetic markers that affect clinical response may result in optimization of anti-
VEGF therapy. An obvious strategy to identify pharmacogenetic markers that may predict antiangiogenic therapy is to analyze polymorphisms in genes within the VEGF signaling pathway.
Objective
• To evaluate the pharmacogenetic relationship between genotypes of single-nucleotide polymorphisms (SNPs) in vascular endothelial growth factor A ( VEGFA ) and its main receptor ( VEGFR2 ) with response to intravitreal injections of ranibizumab or bevacizumab for nAMD.
Copyright restrictions may apply

Study Design
• The Comparison of Age-Related Macular Degeneration Treatments Trials
(CATT) is a multicentered randomized clinical trial to assess the relative safety and efficacy of 2 drugs, ranibizumab and bevacizumab, to treat nAMD. The large cohort of patients treated with anti-VEGF drugs for nAMD in CATT along with the many outcome determinants collected following standardized protocols make this study population an ideal group to evaluate the effects of genetic polymorphisms on treatment response.
Visual acuity (VA) was assessed using eETDRS testing. Optical coherence tomographic (OCT) outcomes were determined by an independent OCT reading center. Fluorescein angiographic (FA) and photographic outcomes were determined by an independent fundus photographic reading center.
Participants
• 835 of 1116 patients (75%) participating in CATT were recruited through 43
CATT clinical centers.
Copyright restrictions may apply

Genotyping
• Each patient was genotyped for 7 SNPs in VEGFA and 1 SNP in the highaffinity VEGFA receptor, VEGFR2 , using TaqMan SNP genotyping assays
(Applied Biosystems).
Data Analysis
• Genotypic frequencies were compared with clinical measures of response to therapy at 1 year including mean VA, mean change in VA,

15-letter increase, retinal thickness, mean change in total foveal thickness, presence of fluid on OCT, presence of leakage on FA, mean change in lesion size, and mean number of injections administered. Differences in response by genotype were evaluated with tests of linear trend calculated from logistic regression models for categorical outcomes and linear regression models for continuous outcomes. The method of controlling the false discovery rates was used to adjust for multiple comparisons.
Copyright restrictions may apply

• The SNPs chosen for evaluation are listed. They represent SNPs within the
VEGF pathway that have been studied for their role in the pathogenesis of
AMD. Several of these SNPs have also been investigated in small-scale studies for their influence on anti-VEGF treatment outcomes. The results to date have been conflicting and inconclusive.
Copyright restrictions may apply

Copyright restrictions may apply
Genotypic Associations With Visual Outcome
Measures and the Number of Injections at 1
Year Among 835 Patients a a The total number of patients with data available for each outcome varied because of missed visits or poor image quality. The percentage missing for each outcome was 5% or less. P values are for linear trend. The genotype is coded as 2 for 2 risk alleles, 1 for
1 risk allele, or 0 for no risk allele. For continuous outcomes, the linear trend P value is from linear regression with the genotype as a continuous variable. For categorical outcomes, the linear trend P value is from logistic regression with the genotype as a continuous variable (cumulative logit model for 3-level outcome).
b The total number of patients with data available for each outcome varied because of missed visits or poor image quality. The percentage missing for each outcome was 5% or less.
• For each of the 3 vision measures evaluated at 1 year, there was no association with any of the genotypes or with the number of risk alleles from the 8 SNPs.

• For each of the 5 anatomical measures, there were few noteworthy associations. Four of the VEGFA SNPs demonstrated an association with retinal thickness (rs699947, rs833070, rs1413711, P = .03 to .04; rs2146323, P = .006). However, the adjusted P values for these associations were all not statistically significant (adjusted P value range,
P = .24 to .45). Furthermore, none of these modest associations were supported by any other anatomical measure.
Copyright restrictions may apply

Copyright restrictions may apply
Genotypic Associations With
Anatomical Outcome Measures at
1 Year Among 835 Patients a a P values are for linear trend. The genotype is coded as 2 for
2 risk alleles, 1 for 1 risk allele, and 0 for no risk allele. For continuous outcomes, the linear trend P value is from linear regression with the genotype as a continuous variable. For categorical outcomes, the linear trend P value is from logistic regression with the genotype as a continuous variable
(cumulative logit model for 3-level outcome).
b The total number of patients with data available for each outcome varied because of missed visits or poor image quality. The percentage missing for each outcome was 5% or less except for 10% missing for lesion size change from baseline.

Copyright restrictions may apply
Genotypic Associations With Visual Outcome
Measures and the Number of Injections at 1 Year
Among 835 Patients a a The total number of patients with data available for each outcome varied because of missed visits or poor image quality. The percentage missing for each outcome was 5% or less. P values are for linear trend. The genotype is coded as 2 for 2 risk alleles, 1 for 1 risk allele, or 0 for no risk allele. For continuous outcomes, the linear trend P value is from linear regression with the genotype as a continuous variable. For categorical outcomes, the linear trend P value is from logistic regression with the genotype as a continuous variable (cumulative logit model for 3-level outcome).
b The total number of patients with data available for each outcome varied because of missed visits or poor image quality. The percentage missing for each outcome was 5% or less.
c One hundred patients are excluded because of having 3 or more missed visits or ever having no treatment because of contraindication or futility in 1 year.
• Among the participants in the 2 as-needed groups, no association was found in the number of injections among the different genotypes for any of the 8 SNPs or for the total number of risk alleles from the 8 SNPs.

• Largest pharmacogenetic analysis exploring the relationship between genotypes and response to anti-VEGF therapy with >90% power to detect a 1-line difference between the highest- and lowest-risk groups.
• The relationship between genotype and response to therapy did not vary by drug or dosing regimen.
• Although specific risk alleles for VEGFA and VEGFR2 may be associated with the development of AMD in some patient populations, they did not predict response to anti-VEGF therapy in CATT study participants.
• Additional studies, including a genome-wide analysis, are under way to identify novel polymorphisms that may be associated with response to anti-
VEGF therapy in patients with nAMD.
Copyright restrictions may apply

• If you have questions, please contact the corresponding author:
– Stephanie A. Hagstrom, PhD, Cole Eye Institute, Cleveland Clinic, 9500
Euclid Ave, Cleveland, OH 44195 (hagstrs@ccf.org).
•
This study was supported by cooperative agreements U10 EY017823, U10
EY017825, U10 EY017826, and U10 EY017828 from the National Eye
Institute.
• None reported.
Copyright restrictions may apply